Image Compression Algorithm Profiling for Image Complexity Estimation

In this project we will experimentally compare several methods of profiling image compression and decompression. This work is inspired by "Image Characterization and Classification by Physical Complexity" by Hector Zenil, Jean-Paul Delahaye, and Cedric Gaucherel



Name of research group, project, or lab
Prof. Bang's ALPAQA Lab
Logistics Information:
Project categories
Computer Science
Signal Processing
Student ranks applicable
Student qualifications

Important interests, skills, or background: 

CS70-level comfort with C++ and Valgrind

image representations (e.g. RGB arrays)

data analysis and plotting with at least one language like Matlab / Octave / Python (numpy, matplotlib, etc.) / Mathematica.


Bonus interest, skills, or background: 

simple compression algorithms (e.g. Huffman coding)

basic information theory (Shannon entropy)


What to put in your application:

Please explain any specific experience with the areas mentioned above and any specific interests in image processing / image complexity. 

Time commitment
Summer - Full Time
Paid Research
Number of openings
Techniques learned

Main software tools to be used are

valgrind, perf, pin, pngcrush, lodepng, picopng, plus analysis and plotting language from above


Main mathematical / algorithmic tools to be used are

Huffman coding, LZ77 compression, traditional and algorithmic information theory, Kolmogorov complexity 

Contact Information:
Mentor name
Lucas Bang
Mentor email
Mentor position
Computer Science Professor
Name of project director or principal investigator
Lucas Bang
Email address of project director or principal investigator
2 sp. | 16 appl.
Hours per week
Summer - Full Time
Project categories
Mathematics (+3)
Computer ScienceMathematicsAlgorithmsSignal Processing